Multimedia resource recommendation methods and related devices

Abstract

This application discloses a multimedia resource recommendation method and related apparatus to solve the problems of complex, inefficient, and inaccurate operation in related technologies for multimedia resource recommendation. This application uses the user's recent historical behavior sequence to mine the user's expected recommendation level for different multimedia resource types. The higher the expected recommendation level, the more the user expects to obtain that type of multimedia resource. Therefore, this application can construct the user's target historical behavior sequence based on the type of multimedia resource the user expects to obtain. This target historical behavior sequence, compared to the long-term behavior sequence in related technologies, focuses more on the user's recent interests and the type of multimedia resource the user recently expects to obtain. Furthermore, this application does not require obtaining the target historical behavior sequence separately for each resource type; it only needs to calculate the target historical behavior sequence once per request. Therefore, the recommendation method of this application simplifies operation and improves recommendation efficiency compared to existing technologies.

Description

Technical Field

[0001] This application relates to the field of multimedia information processing technology, and in particular to a multimedia resource recommendation method and related apparatus. Background Technology

[0002] In the training of short video recommendation system models, information on user interests contained in users' historical behavior can be used to recommend more interesting content to users, which is of guiding significance for the model's learning.

[0003] In related technologies, after obtaining multimedia resources to be recommended based on user requests, all historical user behavior data is used as input to a search-based user behavior interest CTR model (SIM). If the candidate multimedia resources to be recommended include multiple types of multimedia resources, the GSU (General Search Unit) module generates corresponding long-term behavior sequences for each type of multimedia resource to be recommended, which are then used as input to the ESU (Exact Search Unit) module. A deep learning model based on an attention mechanism is then used to model the long-term behavior sequences to obtain evaluation scores for each candidate multimedia resource. Finally, the multimedia resources to be recommended are ranked and recommended based on these evaluation scores. However, the methods for recommending multimedia resources in related technologies are complex, inefficient, and have poor accuracy. Summary of the Invention

[0004] This application provides a multimedia resource recommendation method and related apparatus to solve the problems of complex, inefficient and inaccurate operation when recommending multimedia resources in related technologies.

[0005] Firstly, this application provides a multimedia resource recommendation method, the method comprising:

[0006] In response to a multimedia resource recommendation request from a target account, the recent historical behavior sequence of the target account is obtained; the historical behavior in the recent historical behavior sequence represents the target account's operation behavior on the corresponding multimedia resources within a preset time period before the current time;

[0007] Based on the recent historical behavior sequence, the expected recommendation level of the target account for multiple multimedia resource types in the first multimedia resource type set is determined; wherein, the first multimedia resource type set is obtained based on the multimedia resources corresponding to the historical behaviors in the recent historical behavior sequence;

[0008] Based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set, the target account's interest in multimedia resource types is determined.

[0009] Target historical behaviors that match the type of multimedia resources of interest are filtered out from the full historical behavior sequence of the target account to obtain the target historical behavior sequence corresponding to the target account;

[0010] Based on the target's historical behavior sequence, the target account's level of interest in the multimedia resources to be recommended is determined, and a target multimedia resource among the multimedia resources to be recommended is determined based on the level of interest; the target multimedia resource is used to recommend to the target account.

[0011] In one possible implementation, before determining the target account's expected recommendation level for each of the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence, the method further includes:

[0012] Obtain the multimedia resources corresponding to each historical behavior in the recent historical behavior sequence to obtain the first multimedia resource set;

[0013] The first multimedia resource set is classified to obtain a second multimedia resource type set;

[0014] Determine the target account's level of interest in each media resource type within the second multimedia resource type set;

[0015] The first multimedia resource type set is obtained by filtering out multimedia resource types with interest levels lower than a preset interest level threshold from the second multimedia resource type set.

[0016] In one possible implementation, determining the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence specifically includes:

[0017] Perform the following for each multimedia resource type in the first set of multimedia resource types:

[0018] Determine the number n of multimedia resources among the multimedia resources included in the multimedia resource type, where the access duration of the historical behavior is higher than the duration threshold; where n is a positive integer greater than or equal to 1;

[0019] The expected recommendation level for the multimedia resource type is determined by adopting a relationship that the expected recommendation level is directly proportional to n and inversely proportional to the number of multimedia resources included in the multimedia resource type.

[0020] In one possible implementation, the multimedia resource type includes the following multimedia resources:

[0021] The accessed multimedia resources belonging to the multimedia resource type in the multimedia resources corresponding to the historical behavior of the recent historical behavior sequence.

[0022] And / or,

[0023] The associated multimedia resources of the accessed multimedia resources are multimedia resources that are simultaneously recommended to the target account when the accessed multimedia resources are recommended, and the number of accessed multimedia resources and the total number of associated multimedia resources do not exceed the upper limit.

[0024] In one possible implementation, before determining the target account's level of interest in the recommended multimedia resources based on the target's historical behavior sequence, the method further includes:

[0025] Select candidate multimedia resources belonging to the interest multimedia resource type from the multimedia resources to be recommended;

[0026] If the number of candidate multimedia resources is less than the preset number, then candidate multimedia resources are selected from similar multimedia resource types of the interest multimedia resource type until the total number of finally selected candidate multimedia resources is not less than the preset number.

[0027] The final shortlisted candidate multimedia resources are filtered to obtain the final multimedia resources to be recommended, which are used to determine the target account's level of interest in the multimedia resources to be recommended.

[0028] In one possible implementation, if it is determined that the expected recommendation degree distribution of the plurality of multimedia resource types meets a preset distribution, then the step of determining the target account's interest multimedia resource types based on the target account's expected recommendation degree for the plurality of multimedia resource types is executed.

[0029] In one possible implementation, before determining the target account's preferred multimedia resource types based on the target account's expected recommendation level for each of the multiple multimedia resource types, the method further includes:

[0030] If it is determined that the expected recommendation degree distribution of the multiple multimedia resource types does not meet the preset distribution, then cluster analysis is performed on the multimedia resource types in the first multimedia resource type set to obtain a new first multimedia resource type set, and the process returns to the step of determining the expected recommendation degree of the target account for the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence.

[0031] In one possible implementation, determining the target account's preferred multimedia resource types based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set specifically includes:

[0032] Based on the desired recommendation level from highest to lowest, a specified number of multimedia resource types are selected as the target account's preferred multimedia resource types; or,

[0033] Based on the order of expected recommendation level from high to low, multimedia resource types with expected recommendation level higher than the expected recommendation level threshold are selected as the target account's interest multimedia resource types.

[0034] In one possible implementation, determining the target account's interest in each media resource type within the second multimedia resource type set specifically includes:

[0035] Based on the historical behavior in the recent historical behavior sequence, determine the operation frequency of the target account on each multimedia resource type in the second multimedia resource type set;

[0036] Based on the positive correlation between the interest level and the operation frequency, the interest level of each media resource type in the second multimedia resource type set is determined.

[0037] In one possible implementation, the recent historical behavior sequence includes:

[0038] The target account's specified number of historical behaviors within a preset time period prior to the current time;

[0039] And / or,

[0040] The target account's historical behavior during a preset time period prior to the current time.

[0041] Secondly, this application provides a multimedia resource recommendation device, the device comprising:

[0042] The recent historical behavior sequence acquisition module is configured to execute a multimedia resource recommendation request in response to a target account and acquire the recent historical behavior sequence of the target account; the historical behavior in the recent historical behavior sequence represents the target account's operation behavior on the corresponding multimedia resources within a preset time period before the current time;

[0043] The expected recommendation level determination module is configured to determine the expected recommendation level of the target account for multiple multimedia resource types in a first multimedia resource type set based on the recent historical behavior sequence; wherein, the first multimedia resource type set is obtained based on the multimedia resources corresponding to the historical behaviors in the recent historical behavior sequence;

[0044] The interest multimedia resource type determination module is further configured to determine the interest multimedia resource types of the target account based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set.

[0045] The target historical behavior filtering module is configured to filter out target historical behaviors that match the interest multimedia resource type from the full historical behavior sequence of the target account, and obtain the target historical behavior sequence corresponding to the target account;

[0046] The target multimedia resource determination module is further configured to determine the target account's level of interest in the multimedia resources to be recommended based on the target's historical behavior sequence, and to determine the target multimedia resource among the multimedia resources to be recommended based on the level of interest; the target multimedia resource is used to recommend to the target account.

[0047] In one possible implementation, the device further includes:

[0048] The first multimedia resource set acquisition module is configured to acquire multimedia resources corresponding to each historical behavior in the recent historical behavior sequence before the expected recommendation degree determination module determines the expected recommendation degree of the target account for each of the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence, thereby obtaining the first multimedia resource set.

[0049] The classification module is configured to classify the first multimedia resource set to obtain a second multimedia resource type set.

[0050] The interest determination module is configured to determine the target account's interest in each media resource type in the second multimedia resource type set.

[0051] The first multimedia resource type set determination module is configured to filter out multimedia resource types with interest levels lower than a preset interest level threshold from the second multimedia resource type set to obtain the first multimedia resource type set.

[0052] In one possible implementation, the step of determining the target account's expected recommendation level for multiple multimedia resource types in a first multimedia resource type set based on the recent historical behavior sequence is performed, and the expected recommendation level determination module is specifically configured to perform:

[0053] Perform the following for each multimedia resource type in the first set of multimedia resource types:

[0054] Determine the number n of multimedia resources among the multimedia resources included in the multimedia resource type, where the access duration of the historical behavior is higher than the duration threshold; where n is a positive integer greater than or equal to 1;

[0055] The expected recommendation level for the multimedia resource type is determined by adopting a relationship that the expected recommendation level is directly proportional to n and inversely proportional to the number of multimedia resources included in the multimedia resource type.

[0056] In one possible implementation, the multimedia resource type includes the following multimedia resources:

[0057] The accessed multimedia resources belonging to the multimedia resource type in the multimedia resources corresponding to the historical behavior of the recent historical behavior sequence.

[0058] And / or,

[0059] The associated multimedia resources of the accessed multimedia resources are multimedia resources that are simultaneously recommended to the target account when the accessed multimedia resources are recommended, and the number of accessed multimedia resources and the total number of associated multimedia resources do not exceed the upper limit.

[0060] In one possible implementation, the device further includes:

[0061] The candidate multimedia resource filtering module is configured to filter candidate multimedia resources belonging to the interest multimedia resource type from the multimedia resources to be recommended before the target multimedia resource determination module determines the target account's interest level in the multimedia resources to be recommended based on the target's historical behavior sequence.

[0062] The resource supplementation module is configured to perform the following: if the number of candidate multimedia resources is less than a preset number, then select candidate multimedia resources from similar multimedia resource types of the interest multimedia resource type until the total number of finally selected candidate multimedia resources is not less than the preset number.

[0063] The module for determining multimedia resources to be recommended is configured to perform filtering on the finally selected candidate multimedia resources to obtain the final multimedia resources to be recommended for determining the target account's level of interest in the multimedia resources to be recommended.

[0064] In one possible implementation, the device further includes:

[0065] The expected recommendation degree distribution determination module is configured to determine, before the interest multimedia resource type determination module performs the step of determining the target account's interest multimedia resource types based on the expected recommendation degree of the target account for each of the multiple multimedia resource types, determine that the expected recommendation degree distribution of the multiple multimedia resource types satisfies a preset distribution.

[0066] In one possible implementation, the device further includes:

[0067] The clustering module is configured to perform clustering analysis on the multimedia resource types in the first multimedia resource type set before the interest multimedia resource type determination module determines the interest multimedia resource types of the target account based on the expected recommendation degree of the target account for multiple multimedia resource types. If the expected recommendation degree distribution determination module determines that the expected recommendation degree distribution of the multiple multimedia resource types does not meet the preset distribution, then clustering analysis is performed on the multimedia resource types in the first multimedia resource type set to obtain a new first multimedia resource type set.

[0068] The iteration module is configured to, after obtaining a new first set of multimedia resource types, return to execute the step of determining the target account's expected recommendation level for each of the multiple multimedia resource types in the first set of multimedia resource types based on the recent historical behavior sequence.

[0069] In one possible implementation, the step of determining the target account's preferred multimedia resource types based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set is performed. Specifically, the preferred multimedia resource type determination module is configured to perform:

[0070] Based on the desired recommendation level from highest to lowest, a specified number of multimedia resource types are selected as the target account's preferred multimedia resource types; or,

[0071] Based on the order of expected recommendation level from high to low, multimedia resource types with expected recommendation level higher than the expected recommendation level threshold are selected as the target account's interest multimedia resource types.

[0072] In one possible implementation, the process of determining the target account's interest in each media resource type within the second multimedia resource type set is performed, and the interest determination module is specifically configured to perform:

[0073] Based on the historical behavior in the recent historical behavior sequence, determine the operation frequency of the target account on each multimedia resource type in the second multimedia resource type set;

[0074] Based on the positive correlation between the interest level and the operation frequency, the interest level of each media resource type in the second multimedia resource type set is determined.

[0075] In one possible implementation, the recent historical behavior sequence includes:

[0076] The target account's specified number of historical behaviors within a preset time period prior to the current time;

[0077] And / or,

[0078] The target account's historical behavior during a preset time period prior to the current time.

[0079] Thirdly, this application also provides an electronic device, including:

[0080] processor;

[0081] Memory used to store the processor's executable instructions;

[0082] The processor is configured to execute the instructions to implement any of the multimedia resource recommendation methods provided in the first aspect of this application.

[0083] Fourthly, embodiments of this application also provide a computer-readable storage medium, wherein when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is able to perform any of the multimedia resource recommendation methods provided in the first aspect of this application.

[0084] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements any of the multimedia resource recommendation methods provided in the first aspect of this application.

[0085] The technical solutions provided by the embodiments of this application bring at least the following beneficial effects:

[0086] This application utilizes users' recent historical behavior sequences to mine their expected recommendation levels for different multimedia resource types. A higher expected recommendation level indicates a greater user interest in and desire for that type of multimedia resource. Therefore, this application, based on users' recent historical behavior sequences, can mine users' recent interests and their level of interest in each multimedia resource type. Then, based on the types of multimedia resources the user expects to obtain, a target historical behavior sequence is constructed. This target historical behavior sequence, compared to the long-term behavior sequences of related technologies, places greater emphasis on the user's recent interests and the types of multimedia resources the user recently expects to obtain. Based on this application's target historical behavior sequence, it can better describe the user's needs and make accurate recommendations. Furthermore, this application does not require obtaining a target historical behavior sequence separately for each multimedia resource type; it only needs to calculate the target historical behavior sequence once per request. Therefore, this application's recommendation method simplifies operations and improves recommendation efficiency compared to existing technologies.

[0087] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The purposes and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0088] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0089] Figure 1 This is a schematic diagram illustrating an application scenario of the multimedia resource recommendation method provided in the embodiments of this application;

[0090] Figure 2 A flowchart illustrating the training of a multimedia resource recommendation model provided in an embodiment of this application;

[0091] Figure 3 A flowchart illustrating the multimedia resource recommendation method provided in this application embodiment;

[0092] Figure 4 A flowchart illustrating the method for determining the first multimedia resource type set provided in an embodiment of this application;

[0093] Figure 5A flowchart illustrating a method for determining the desired level of recommendation, provided in an embodiment of this application;

[0094] Figure 6 A flowchart illustrating the method for determining multimedia resources to be recommended, as provided in an embodiment of this application;

[0095] Figure 7 A block diagram of a multimedia resource recommendation device provided in an embodiment of this application;

[0096] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0097] To enable those skilled in the art to better understand the technical solutions of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0098] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data used can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0099] In addition, it should be noted that the acquisition, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations.

[0100] The following explanations of some terms used in the embodiments of this application are provided to facilitate understanding by those skilled in the art.

[0101] (1) In the embodiments of this application, the term "multiple" refers to two or more, and other quantifiers are similar.

[0102] (2) "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three cases: A alone, A and B at the same time, and B alone. The character " / " generally indicates that the related objects before and after it are in an "or" relationship.

[0103] (3) A server is a service for a terminal. The services include providing resources to the terminal and storing terminal data. A server is a counterpart to the application installed on the terminal and works in conjunction with the application on the terminal.

[0104] (4) Terminal devices can refer to both software applications (APPs) and client devices. They have a visual display interface and can interact with the target account; they correspond to a server and provide local services to the client. For software applications, except for some applications that run only locally, they are generally installed on ordinary client terminals and need to cooperate with the server to run. With the development of the Internet, commonly used applications include short video applications, email clients for sending and receiving emails, and instant messaging clients. For these types of applications, corresponding servers and service programs are needed on the network to provide services such as database services and configuration parameter services. Therefore, a specific communication connection needs to be established between the client terminal and the server to ensure the normal operation of the application.

[0105] In the training of short video recommendation system models, information on user interests contained in users' historical behavior can be used to recommend more interesting content to users, which is of guiding significance for the model's learning.

[0106] In related technologies, after obtaining multimedia resources to be recommended based on user requests, all historical behavior data of the user are used as input to a search-based user behavior interest CTR model (SIM). If the candidate multimedia resources to be recommended include multiple types of multimedia resources, the GSU (General Search Unit) module generates corresponding long-term behavior sequences for each type of multimedia resource to be recommended as input to the ESU (Exact Search Unit) module. Then, a deep learning model based on the attention mechanism is used to model the long-term behavior sequences to obtain the evaluation scores of each candidate multimedia resource. Finally, the multimedia resources to be recommended are ranked and recommended based on the evaluation scores.

[0107] Therefore, in related technologies, a single user request requires generating a long-term behavior sequence based on the user's entire historical behavior and all resources to be recommended. Since a single user request may correspond to hundreds or even thousands of candidate videos, the process of building a long-term behavior sequence for a single user request may need to be executed hundreds or thousands of times. Furthermore, each time a long-term behavior sequence is built, the GSU module needs to be executed for each candidate video, resulting in a high request frequency. Consequently, the process of building long-term behavior sequences in related technologies is complex and inefficient, leading to a complex, inefficient, and inaccurate recommendation process.

[0108] Furthermore, the long-term behavior sequences generated by using all of a user's historical behavior in related technologies place more emphasis on the user's long-term historical interests. As a result, the resources filtered based on long-term historical interests are often resources that the user liked in the past, while the resources that the user currently likes cannot be satisfied, resulting in the recommended resources often failing to meet the user's current needs.

[0109] In view of this, this application provides a multimedia resource recommendation method and related apparatus to solve the problems of complex, inefficient and inaccurate operation when recommending multimedia resources in related technologies.

[0110] The inventive concept of this application can be summarized as follows: First, in response to a multimedia resource recommendation request from a target account, the recent historical behavior sequence of the target account is obtained. Then, based on the recent historical behavior sequence, the expected recommendation level of the target account for multiple multimedia resource types in a first multimedia resource type set is determined. Simultaneously, based on the expected recommendation level of the target account for multiple multimedia resource types in the first multimedia resource type set, the target account's interest in multimedia resource types is determined. Target historical behaviors matching the interest in multimedia resource types are then filtered from the target account's full historical behavior sequence to obtain the target historical behavior sequence corresponding to the target account. Finally, based on the target historical behavior sequence, the target account's level of interest in the multimedia resources to be recommended is determined, and based on the level of interest, the target account's interest in the recommended multimedia resources is determined. This application describes the target multimedia resources among the multimedia resources to be recommended and recommends them to the target account. Therefore, this application uses the user's recent historical behavior sequence to mine the user's expected recommendation level for different multimedia resource types. The higher the expected recommendation level, the more interested the user is in that type of multimedia resource and the more they desire to obtain it. Thus, this application can mine the user's recent interests and the user's interest level for each type of multimedia resource based on the user's recent historical behavior sequence. Then, based on the type of multimedia resource the user expects to obtain, a target historical behavior sequence is constructed for the user. This target historical behavior sequence, compared to the long-term behavior sequence of related technologies, focuses more on the user's recent interests and the type of multimedia resource the user recently expects to obtain. Based on the target historical behavior sequence of this application, it can better describe the user's needs and make accurate recommendations. Furthermore, this application does not require obtaining the target historical behavior sequence separately for each multimedia resource type; it only needs to calculate the target historical behavior sequence once per request. Therefore, the recommendation method of this application simplifies the operation and improves recommendation efficiency compared to existing technologies.

[0111] After introducing the design concept of the embodiments of this application, the following is a brief introduction to the application scenarios to which the technical solutions of the embodiments of this application can be applied. It should be noted that the application scenarios described below are only for illustrating the embodiments of this application and are not intended to limit the scope. In specific implementation, the technical solutions provided by the embodiments of this application can be flexibly applied according to actual needs.

[0112] refer to Figure 1 This is a schematic diagram illustrating an application scenario of the multimedia resource recommendation method provided in this application embodiment. The application scenario includes multiple terminal devices 101 (including terminal device 101-1, terminal device 101-2, ..., terminal device 101-n) and a server 102. The terminal devices 101 and the server 102 are connected via a wireless or wired network, and the server 102 provides multimedia resources to the terminal devices 101 for display.

[0113] Terminal devices 101 include, but are not limited to, electronic devices such as desktop computers, mobile phones, mobile computers, tablet computers, media players, smart wearable devices, and smart TVs.

[0114] Server 102 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Server 102 can be an independent physical server, a server cluster consisting of multiple physical servers, or a distributed system. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0115] Of course, the methods provided in the embodiments of this application are not limited to... Figure 1 The application scenarios shown can also be used in other possible application scenarios, and the embodiments of this application do not impose limitations. Figure 1 The functions that each device in the application scenario shown can achieve will be described in subsequent method embodiments, and will not be elaborated on here.

[0116] To further illustrate the technical solutions provided in the embodiments of this application, a detailed description is provided below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiments of this application provide method operation steps as shown in the following embodiments or drawings, the method may include more or fewer operation steps based on conventional or non-inventive methods. In steps where there is no logically necessary causal relationship, the execution order of these steps is not limited to the execution order provided in the embodiments of this application.

[0117] It should be noted that the resource recommendation method provided in this application is applicable to any online resource, such as short videos, long videos, network elements, and products, in scenarios where online resource recommendations are needed. Furthermore, all user information required to achieve resource recommendations is obtained through user authorization.

[0118] The training process of the multimedia resource recommendation model to which the technical solutions of the embodiments of this application are applicable is briefly introduced below, in order to facilitate those skilled in the art to understand the technical solutions provided by the embodiments of this application.

[0119] refer to Figure 2 The diagram below illustrates the training process of the multimedia resource recommendation model provided in this application embodiment, including the following steps:

[0120] In step 201, training samples are obtained to construct the multimedia resource recommendation model (i.e., the user behavior interest CTR model mentioned above).

[0121] In some embodiments, the training samples of the multimedia resource recommendation model include target account features, multimedia resource features, contextual features, and the target account's behavior toward the sample, such as liking, following, or watching the sample for a long time. The training samples also include the user's target historical behavior sequence. In this application, the target historical behavior sequence is constructed based on the user's recent historical behavior sequence (which will be explained later).

[0122] The target account features can be the target account ID, device ID, or other features that characterize the target account information, such as the target account's interests or age. Target account features can also be an average summation of target account behavior sequences, such as an average summation of the IDs of recently viewed multimedia resources. Furthermore, target account features can be target account behavior sequences, such as a sequence of target account historical viewed multimedia resource IDs, a sequence of target account historical viewed multimedia resource author IDs, a sequence of target account historical viewed multimedia resource durations, or a sequence of target account historical viewed multimedia resources from the current time. These can be set according to actual usage, and this application embodiment does not limit this.

[0123] The multimedia resource characteristics can be multimedia resource IDs or other characteristics that can characterize multimedia resource information, such as multimedia resource age, multimedia resource category, multimedia resource type, etc., which can be set according to actual usage. This application embodiment does not limit this.

[0124] In step 202, the structure and parameters of the ESU in the multimedia resource recommendation model are adjusted based on the training samples, i.e., the ESU module is trained. Specifically, this can be implemented by modeling the target historical behavior sequence obtained in step 201 based on an attention mechanism.

[0125] In step 203, the GSU module in the multimedia resource recommendation model is trained, thereby obtaining the trained GSU module and ESU module.

[0126] In step 204, multimedia resource recommendations are made based on the trained multimedia resource model.

[0127] The multimedia resource recommendation method proposed in this application is mainly based on the above-mentioned process of training a multimedia resource recommendation model. It uses the user's recent historical behavior sequence to mine the user's expected recommendation level for different multimedia resource types, and obtains the multimedia resource types the user is currently interested in based on the user's expected recommendation level for different multimedia resource types, thereby obtaining the user's sample historical behavior sequence, i.e., obtaining the training samples in step 201. After obtaining the training samples, steps 202 and 203 are used to train the multimedia resource recommendation model. The multimedia resource recommendation method provided in this application mainly uses the same method as obtaining the sample historical behavior sequence to obtain the user's target historical behavior sequence, and inputs the target historical behavior sequence into the trained multimedia resource recommendation model, thereby obtaining target multimedia resources to recommend to the user.

[0128] Reference Figure 3 This is a flowchart illustrating a multimedia resource recommendation method provided in an embodiment of this application. Figure 3 As shown, this method can be implemented as follows:

[0129] In step 301, in response to the multimedia resource recommendation request of the target account, the recent historical behavior sequence of the target account is obtained; the historical behavior in the recent historical behavior sequence represents the target account's operation behavior on the corresponding multimedia resources within a preset time period before the current time.

[0130] For example, if the preset time period is set to 10 minutes, then the historical behavior in the recent historical behavior sequence represents the target account's operation behavior on the corresponding multimedia resources within 10 minutes before the current time. For example, the recent historical behavior sequence includes the viewing time or number of likes for each multimedia resource within 10 minutes.

[0131] In some embodiments, a multimedia resource recommendation request may be a refresh operation of the target account on multimedia resources or a search operation of the target account on multimedia resources; this application does not specifically limit this.

[0132] In some embodiments, the target account may include a unique identifier for the target account, such as a target account ID, and may also include a device ID, which can be set according to actual usage. This application embodiment does not limit this.

[0133] In some embodiments, in response to a multimedia resource recommendation request from a target account, resources in the multimedia resource library are filtered through steps such as vector recall and coarse ranking to ultimately obtain multimedia resources to be recommended. The number of multimedia resources to be recommended is relatively large; for example, it can typically exceed 1000.

[0134] In some embodiments, the recent historical behavior sequence includes: a specified number of historical behaviors of the target account within a preset time period prior to the current time; and / or, the historical behaviors of the target account within a preset time period prior to the current time. For example, taking multimedia resources as videos, the recent historical behavior sequence may include a sequence of video IDs of the target account's most recently viewed videos, a sequence of video author IDs of the target account's most recently viewed videos, a sequence of video durations of the target account's most recently viewed videos, a sequence of the length of time since the target account's most recently viewed videos, etc. It may also include a sequence of video IDs of videos watched by the target account within the last 10 minutes, or the number of times the target account liked the videos watched within the last 10 minutes.

[0135] Therefore, by setting a time period or specifying a quantity, a reliable range can be determined for the recent historical behavior series obtained from multimedia resource recommendation requests based on the target account. While obtaining historical behavior, limiting the quantity can reduce the processing volume of historical behavior and also reflect the user's recent preferences.

[0136] In response to a multimedia resource recommendation request from a target account, the step of obtaining the target account's recent historical behavior sequence requires obtaining information about the target account. Therefore, in this application, any information about the target account is obtained with authorization and consent.

[0137] In step 302, based on the recent historical behavior sequence, the expected recommendation level of the target account for multiple multimedia resource types in the first multimedia resource type set is determined; wherein, the first multimedia resource type set is obtained based on the multimedia resources corresponding to the historical behaviors in the recent historical behavior sequence.

[0138] The expected recommendation level refers to the degree to which the target account's interest in a certain type of multimedia resource is not satisfied.

[0139] In one possible implementation, determining the expected recommendation level for all multimedia resource types in a recent behavior sequence may take a considerable amount of time, making the determination of the expected recommendation level inefficient. Therefore, to improve the efficiency of determining the expected recommendation level, in this embodiment of the application, before determining the expected recommendation level of the target account for each of the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence, the range of multimedia resource types can be narrowed down. Specifically, this can be done as follows: Figure 4 The steps shown are as follows:

[0140] In step 401, the multimedia resources corresponding to each historical behavior in the recent historical behavior sequence are obtained to obtain the first multimedia resource set.

[0141] In step 402, the first multimedia resource set is classified to obtain the second multimedia resource type set.

[0142] In step 403, the target account’s interest in each media resource type in the second multimedia resource type set is determined.

[0143] In one possible implementation, the target account's interest in each media resource type in the second multimedia resource type set is determined in this embodiment of the application. Specifically, it can be implemented as follows: first, based on the historical behavior in the recent historical behavior sequence, the operation frequency of the target account on each multimedia resource type in the second multimedia resource type set is determined; then, based on the positive correlation between interest and operation frequency, the interest in each media resource type in the second multimedia resource type set is determined.

[0144] For example, based on historical behaviors in recent historical behavior sequences, the second set of multimedia resource types is determined to include game videos, food videos, food advertisements, and sales videos. Among these, there are 20 recommended game videos, 15 food videos, 3 food advertisements, and 5 sales videos. The target account liked 10 game videos, 15 food videos, 0 food advertisements, and 1 sales video. Therefore, the target account's operation frequency for game videos is 0.5, for food videos is 1, for food advertisements is 0, and for sales videos is 0.2. Since interest is positively correlated with operation frequency (i.e., the higher the operation frequency, the higher the interest in the corresponding multimedia resource type), the target account's interest in game videos is determined to be 0.5, for food videos is 1, for food advertisements is 0, and for sales videos is 0.2.

[0145] Therefore, by determining the frequency of operations on various types of multimedia resources in recent historical behavior sequences, we can mine users' interest in different types of multimedia resources from user behavior and accurately measure users' interest in different types of resources.

[0146] In step 404, multimedia resource types with interest levels lower than a preset interest level threshold are filtered out from the second multimedia resource type set to obtain the first multimedia resource type set.

[0147] For example, if the interest threshold is set to 0.5, and in the example above, the target account has an interest of 0.5 in game videos, 1 in food videos, 0 in food advertisements, and 0.2 in videos selling goods, then multimedia resource types with an interest of less than 0.5 in the second multimedia resource type set can be filtered out, resulting in game videos and food videos, which are then used as multimedia resource types in the first multimedia resource type set.

[0148] Therefore, by setting an interest threshold in advance, multimedia resource types that exceed the interest threshold can be filtered out, thereby narrowing down the range of multimedia resource types and improving the efficiency of determining the expected recommendation level of a target account for multiple multimedia resource types in the first multimedia resource type set.

[0149] In one possible implementation, after determining the first set of multimedia resource types, it is necessary to determine the target account's desired recommendation level for each of the multiple multimedia resource types in the first set of multimedia resource types based on the target account's recent historical behavior sequence. In this embodiment, the following steps can be performed separately for each multimedia resource type in the first set of multimedia resource types. Figure 5 The steps shown are as follows:

[0150] In step 501, the number n of multimedia resources whose access duration for historical behavior exceeds a duration threshold is determined among the multimedia resources included in the multimedia resource type; where n is a positive integer greater than or equal to 1.

[0151] In one possible implementation, the multimedia resource type in this application embodiment includes, specifically, the multimedia resources that belong to the multimedia resource type among the multimedia resources corresponding to the historical behavior of the recent historical behavior sequence. Thus, the multimedia resources provided by the user's historical behavior operation, i.e., the multimedia resources that have been accessed, can better represent the resources that the user is interested in.

[0152] In another embodiment, the multimedia resource type may also include associated multimedia resources of the aforementioned accessed multimedia resources. These associated multimedia resources are those simultaneously recommended to the target account when recommending accessed multimedia resources, and the total number of accessed multimedia resources and associated multimedia resources does not exceed the upper limit. That is, when there is insufficient accessed multimedia resource type, associated multimedia resources can be used to supplement it, thereby expanding the number of multimedia resources and ensuring that there are sufficient multimedia resources to support and improve recommendation efficiency when making subsequent recommendations based on multimedia resources.

[0153] In step 502, the expected recommendation level for a multimedia resource type is determined by adopting a relationship that the expected recommendation level is directly proportional to n and inversely proportional to the number of multimedia resources included in the multimedia resource type.

[0154] In one possible implementation, the first multimedia resource type set includes multiple multimedia resource types. Therefore, to calculate the expected recommendation level for each multimedia resource type, the historical behavior in the recent behavior sequence of the target account can be statistically analyzed. The historical behavior of the target account towards each multimedia resource type is divided into positive samples and negative samples. The expected recommendation level of the target account for each multimedia resource type is calculated as the number of positive samples / (number of positive samples + number of negative samples). The higher the proportion of positive samples, the higher the expected recommendation level of the target account for the corresponding multimedia resource type, and the greater the demand of the target account for that multimedia resource type.

[0155] For example, if the multimedia resource types in the first multimedia resource type set include game videos and food videos, and 10 game videos and 5 food videos were recently recommended to the target account, and the target account only recently accessed 6 of the game videos and 1 of the food videos, and the access duration of 1 game video and 1 food video both exceeded the duration threshold.

[0156] The first hypothesis is that the multimedia resource type includes accessed multimedia resources belonging to the multimedia resource type corresponding to the recent historical behavior sequence. In this case, the number of multimedia resources for game videos is 6, that is, the total number of samples is 6, the number of positive samples is 1, the number of negative samples is 5, and the calculated expected recommendation degree is 0.2. On the other hand, the number of multimedia resources for food videos is 1, that is, the total number of samples is 1, the number of positive samples is 1, the number of negative samples is 0, and the calculated expected recommendation degree is 1. Obviously, the target account has a higher expected recommendation degree for multimedia resources of the food video type.

[0157] The second hypothesis is that the multimedia resource type includes accessed multimedia resources belonging to the multimedia resource type and associated multimedia resources of the accessed multimedia resources, which are part of the recent historical behavior sequence. Therefore, the number of multimedia resources for game videos is 10, meaning the total sample size is 10, with 1 positive sample and 9 negative samples, resulting in an expected recommendation level of 0.1. Conversely, the number of multimedia resources for food videos is 5, meaning the total sample size is 5, with 1 positive sample and 4 negative samples, resulting in an expected recommendation level of 0.2. Clearly, the target account has a higher expected recommendation level for multimedia resources of the food video type.

[0158] Therefore, by determining the target account's desired recommendation level for each multimedia resource type in the first multimedia resource type set using the above method, it is possible to identify which types of multimedia resources the user desires, thereby improving the accuracy of the recommendation.

[0159] Therefore, the degree to which the target account's interest in each multimedia resource type is not satisfied can be determined by the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set, and the recommended multimedia resource types can be guaranteed to match the multimedia resource types that the target account has been interested in recently, thus solving the problem of target account interest changes in target historical behavior sequence modeling.

[0160] In step 303, the target account's preferred multimedia resource types are determined based on the target account's expected recommendation level for each of the multiple multimedia resource types in the first multimedia resource type set.

[0161] In one possible implementation, there may not be a target account with a prominent expected recommendation level for a certain multimedia resource type, i.e., there is no prominent interest. This indicates that the target account's most desired multimedia resource type and the least desired multimedia resource type have equal expected recommendation levels. Therefore, before determining the target account's interested multimedia resource types based on the target account's expected recommendation levels for multiple multimedia resource types in the first multimedia resource type set, this embodiment of the application also needs to determine that the expected recommendation level distribution of multiple multimedia resource types satisfies a preset distribution, i.e., it is determined that there is a prominent interest.

[0162] The preset distribution can be represented by the difference between the expected recommendation levels. A gap threshold can be set; if the difference between the expected recommendation levels of multiple multimedia resource types is greater than the gap threshold, it indicates that the distribution of the expected recommendation levels of the multiple multimedia resource types meets the preset distribution. Alternatively, the distribution can be represented as concentrated or dispersed, with a concentration threshold or a dispersion threshold set. If the distribution of the expected recommendation levels of multiple multimedia resource types is greater than the concentration threshold or less than the dispersion threshold, it indicates that the distribution of the expected recommendation levels of the multiple multimedia resource types meets the preset distribution.

[0163] In one possible implementation, if it is determined that the expected recommendation degree distribution of multiple multimedia resource types satisfies a preset distribution, it indicates that there is a target account whose expected recommendation degree for at least one multimedia resource type is greater than the expected recommendation degree for the other multimedia resource types. Then, the step of determining the target account's interest multimedia resource types based on the target account's expected recommendation degree for each of the multiple multimedia resource types is executed.

[0164] Therefore, by setting a preset distribution, the expected recommendation level of multimedia resource types can be determined to meet the requirements for determining the target account's interest in multimedia resource types, thereby better determining the target account's interest in multimedia resource types.

[0165] In one possible implementation, if the distribution of expected recommendation levels for multiple multimedia resource types does not meet a preset distribution, it indicates that there may not be a target account with a particularly high expected recommendation level for a certain multimedia resource type. In this case, cluster analysis is performed on the multimedia resource types in the first multimedia resource type set to obtain a new first multimedia resource type set. The process then returns to the step of determining the expected recommendation level of the target account for each of the multiple multimedia resource types in the first multimedia resource type set based on recent historical behavior sequences. For example, if the expected recommendation levels of 1000 multimedia resource types do not meet the preset distribution, cluster analysis is performed on the 1000 multimedia resource types, reducing them to 100 multimedia resource types. Then, the expected recommendation level of each multimedia resource type is determined based on these 100 types, thereby selecting multiple multimedia resource types as the target account's preferred multimedia resource types.

[0166] In one possible implementation, cluster analysis can be based on the tree structure of hierarchical clustering to merge multimedia resource types with adjacent expected recommendation levels into a single multimedia resource type.

[0167] Therefore, by performing cluster analysis on multimedia resource types that do not meet the preset distribution, the degree of subdivision of multimedia resource types can be reduced, so that the distribution of multimedia resource types in recent behavior sequences is as concentrated as possible, and the difference in the expected recommendation degree of different multimedia resource types is as large as possible, so as to better calculate the expected recommendation degree of multimedia resource types, and better determine the target account's interest in multimedia resource types based on the expected recommendation degree of multimedia resource types.

[0168] In one possible implementation, after determining the target account's expected recommendation level for each multimedia resource type in the first multimedia resource type set, this embodiment of the application can select a specified number of multimedia resource types as the target account's interest multimedia resource types in descending order of expected recommendation level; or, in descending order of expected recommendation level, select multimedia resource types with expected recommendation levels higher than the expected recommendation level threshold as the target account's interest multimedia resource types.

[0169] For example, 100 multimedia resource types can be sorted from highest to lowest according to their expected recommendation level. If a specified number of interest-based multimedia resource types is set to 10, then the top 10 multimedia resource types in the sorted list will be selected as the target account's interest-based multimedia resource types. Alternatively, an expected recommendation level threshold of 0.6 can be set, and multimedia resource types with an expected recommendation level greater than or equal to the threshold of 0.6 will be selected as the target account's interest-based multimedia resource types.

[0170] Therefore, the types of multimedia resources that a target account is interested in can be filtered out by sorting or setting thresholds. This method can filter out the types of multimedia resources that users are interested in, and the filtering method is simple, efficient and easy to implement.

[0171] In step 304, target historical behaviors that match the type of multimedia resources of interest are filtered out from the full historical behavior sequence of the target account to obtain the target historical behavior sequence corresponding to the target account.

[0172] The full historical behavior sequence includes all behavior sequences of the target account since it registered with the application. The sequence length varies for different target accounts, ranging from several hundred to hundreds of thousands. The full historical behavior sequence may include the ID sequence of all multimedia resources accessed by the target account, the author ID sequence of all multimedia resources accessed by the target account, the duration sequence of all multimedia resources accessed by the target account, and the sequence of all multimedia resources accessed by the target account based on the current time. These settings can be configured according to actual usage, and this application embodiment does not limit them.

[0173] In one possible implementation, a new target historical behavior sequence can be constructed by obtaining the corresponding behavior sequence from the full historical behavior sequence of the target account based on the target account's interest in multimedia resources obtained in step 303.

[0174] In step 305, based on the target's historical behavior sequence, the target account's level of interest in the multimedia resources to be recommended is determined, and the target multimedia resources among the multimedia resources to be recommended are determined based on the level of interest; the target multimedia resources are used to recommend to the target account.

[0175] In one possible implementation, before determining the target account's level of interest in the recommended multimedia resources based on the target's historical behavior sequence, this embodiment of the application may further determine the multimedia resources to be recommended, which may specifically involve executing the following steps: Figure 6 The steps shown are as follows:

[0176] In step 601, candidate multimedia resources belonging to the interest multimedia resource type are selected from the multimedia resources to be recommended.

[0177] In step 602, if the number of candidate multimedia resources is less than the preset number, then candidate multimedia resources are selected from similar multimedia resource types of interest multimedia resource types until the total number of finally selected candidate multimedia resources is not less than the preset number.

[0178] In one possible implementation, multiple candidate multimedia resources corresponding to the target account's interest multimedia resource types obtained in step 303 can be filtered from the multimedia resource library. These candidate multimedia resources can be ranked from newest to oldest, or they can be the most popular multimedia resources. When the number of multimedia resources obtained from the multimedia resource library does not meet a preset number, the requirement can be relaxed, and multimedia resources corresponding to other multimedia resource types closest to the interest multimedia resource type can be filtered until the number of filtered multimedia resources meets the preset number. The preset number can be several thousand or tens of thousands, and can be set according to actual conditions; this embodiment does not impose such limitations.

[0179] In step 603, the final selected candidate multimedia resources are filtered to obtain the final multimedia resources to be recommended, which are used to determine the target account's level of interest in the recommended multimedia resources.

[0180] In one possible implementation, the candidate multimedia resources selected in step 602 are used as one recall source. After coarse and fine sorting steps, a funnel-like screening process is performed to finally select the multimedia resources to be recommended. The number of selected multimedia resources to be recommended is generally around 1000, but this can be set according to actual needs; this embodiment does not impose any limitation on this.

[0181] This ensures a sufficient supply of multimedia resources corresponding to the types of multimedia resources that the target account most desires to be recommended, thus solving the problem of insufficient supply of multimedia resources that the target account is interested in.

[0182] In one possible implementation, the target account's level of interest in the recommended multimedia resources is determined based on the target's historical behavior sequence. In this embodiment, the target's historical behavior sequence can be used to sort the recommended multimedia resources to obtain the recommendation order of the recommended multimedia resources. The higher the recommendation order, the greater the target account's level of interest in the recommended multimedia resources.

[0183] For example, the historical behavior sequence of the sample can be obtained first using the method of obtaining the target historical behavior sequence. Then, the historical behavior sequence of the sample can be modeled based on attention mechanisms such as Transformer (machine translation attention mechanism) or Multi-head Attention. For instance, a QKV (query-key-value, attention mechanism) based attention method can be used, taking the multimedia resource features including the target item and other feature vectors as the query, and the historical behavior sequence of the sample as the key and value, to obtain the feature representation of the historical behavior sequence. This feature is then concatenated with other features and passed through an MLP (Multi-Layer Perceptron) to obtain the predicted output of the corresponding target item. Then, the multimedia resource recommendation model is trained, that is, to predict the score of the target account's behavior towards the recommended multimedia resource sample. The target account features, multimedia resource features, context features, and the historical behavior sequence of the sample are input into the model to predict the score of the target account's behavior towards the recommended multimedia resource sample, such as the probability of liking, following, or watching the recommended multimedia resource sample for a long time. The parameters of the multimedia resource recommendation model are updated by calculating the loss function based on the predicted scores of the target account's behavior towards the multimedia resource samples to be recommended and the actual behavior of the target account in the neural network training samples. Finally, based on the ensemble sort formula, the trained multimedia resource recommendation model is used to combine the predicted scores of the target account's various behaviors towards the multimedia resource samples to be recommended, thereby obtaining the recommendation order of the multimedia resources to be recommended and finally determining the target account's level of interest in the multimedia resources to be recommended.

[0184] In one possible implementation, target multimedia resources are determined from the multimedia resources to be recommended based on their level of interest; these target multimedia resources are then recommended to the target account. In this embodiment, the multimedia resources with higher recommendation order are recommended to the target account, meaning those resources that the target account shows greater interest in are recommended to the target account.

[0185] Based on the foregoing description, this embodiment first responds to a multimedia resource recommendation request from a target account by obtaining the target account's recent historical behavior sequence. Then, based on the recent historical behavior sequence, it determines the target account's desired recommendation level for multiple multimedia resource types in the first multimedia resource type set. Simultaneously, based on the target account's desired recommendation level for multiple multimedia resource types in the first multimedia resource type set, it determines the target account's interest in multimedia resource types. Furthermore, it filters out target historical behaviors matching the interest in multimedia resource types from the target account's full historical behavior sequence to obtain the target historical behavior sequence corresponding to the target account. Finally, based on the target historical behavior sequence, it determines the target account's level of interest in the multimedia resources to be recommended, identifies target multimedia resources among the recommended multimedia resources based on the level of interest, and recommends the target multimedia resources to the target account.

[0186] Therefore, this application uses the user's recent historical behavior sequence to mine the user's expected recommendation level for different multimedia resource types. The higher the expected recommendation level, the more interested the user is in that type of multimedia resource and the more they expect to obtain it. Thus, this application can mine the user's recent interests and the user's interest level for each type of multimedia resource based on the user's recent historical behavior sequence. Then, based on the type of multimedia resource the user expects to obtain, a target historical behavior sequence is constructed for the user. Compared with the long-term behavior sequence of related technologies, the target historical behavior sequence focuses more on the user's recent interests and the type of multimedia resource the user recently expects to obtain. Based on the target historical behavior sequence of this application, it can better describe the user's needs and make accurate recommendations for the user. In addition, this application does not need to obtain the target historical behavior sequence for each type of multimedia resource separately. The target historical behavior sequence is calculated once for each request. Therefore, the recommendation method of this application can simplify the operation and improve the recommendation efficiency compared with the prior art.

[0187] Based on the same inventive concept, embodiments of this application also provide a multimedia resource recommendation device. Figure 7 A block diagram of a multimedia resource recommendation device provided in an embodiment of this application, referring to... Figure 7 The device includes: a recent historical behavior sequence acquisition module 701, a desired recommendation level determination module 702, an interest multimedia resource type determination module 703, a target historical behavior filtering module 704, and a target multimedia resource determination module 705, wherein:

[0188] The recent historical behavior sequence acquisition module 701 is configured to execute a multimedia resource recommendation request in response to a target account and acquire the recent historical behavior sequence of the target account; the historical behavior in the recent historical behavior sequence represents the operation behavior of the target account on the corresponding multimedia resources within a preset time period before the current time;

[0189] The expected recommendation level determination module 702 is configured to determine the expected recommendation level of the target account for multiple multimedia resource types in a first multimedia resource type set based on the recent historical behavior sequence; wherein, the first multimedia resource type set is obtained based on the multimedia resources corresponding to the historical behaviors in the recent historical behavior sequence;

[0190] The interest multimedia resource type determination module 703 is further configured to determine the interest multimedia resource types of the target account based on the expected recommendation level of the target account for multiple multimedia resource types in the first multimedia resource type set.

[0191] The target historical behavior filtering module 704 is configured to filter out target historical behaviors that match the interest multimedia resource type from the full historical behavior sequence of the target account, and obtain the target historical behavior sequence corresponding to the target account;

[0192] The target multimedia resource determination module 705 is further configured to determine the target account's level of interest in the multimedia resources to be recommended based on the target's historical behavior sequence, and determine the target multimedia resource among the multimedia resources to be recommended based on the level of interest; the target multimedia resource is used to recommend to the target account.

[0193] In one possible implementation, the device further includes:

[0194] The first multimedia resource set acquisition module is configured to acquire multimedia resources corresponding to each historical behavior in the recent historical behavior sequence before the expected recommendation degree determination module 702 determines the expected recommendation degree of the target account for each of the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence, thereby obtaining the first multimedia resource set.

[0195] The classification module is configured to classify the first multimedia resource set to obtain a second multimedia resource type set.

[0196] The interest determination module is configured to determine the target account's interest in each media resource type in the second multimedia resource type set.

[0197] The first multimedia resource type set determination module is configured to filter out multimedia resource types with interest levels lower than a preset interest level threshold from the second multimedia resource type set to obtain the first multimedia resource type set.

[0198] In one possible implementation, the step of determining the expected recommendation level of the target account for multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence is performed, and the expected recommendation level determination module 702 is specifically configured to perform:

[0199] Perform the following for each multimedia resource type in the first set of multimedia resource types:

[0200] Determine the number n of multimedia resources among the multimedia resources included in the multimedia resource type, where the access duration of the historical behavior is higher than the duration threshold; where n is a positive integer greater than or equal to 1;

[0201] The expected recommendation level for the multimedia resource type is determined by adopting a relationship that the expected recommendation level is directly proportional to n and inversely proportional to the number of multimedia resources included in the multimedia resource type.

[0202] In one possible implementation, the multimedia resource type includes the following multimedia resources:

[0203] The accessed multimedia resources belonging to the multimedia resource type in the multimedia resources corresponding to the historical behavior of the recent historical behavior sequence.

[0204] And / or,

[0205] The associated multimedia resources of the accessed multimedia resources are multimedia resources that are simultaneously recommended to the target account when the accessed multimedia resources are recommended, and the number of accessed multimedia resources and the total number of associated multimedia resources do not exceed the upper limit.

[0206] In one possible implementation, the device further includes:

[0207] The candidate multimedia resource filtering module is configured to filter candidate multimedia resources belonging to the interest multimedia resource type from the multimedia resources to be recommended before the target multimedia resource determination module 705 determines the target account's interest level in the multimedia resources to be recommended based on the target's historical behavior sequence.

[0208] The resource supplementation module is configured to perform the following: if the number of candidate multimedia resources is less than a preset number, then select candidate multimedia resources from similar multimedia resource types of the interest multimedia resource type until the total number of finally selected candidate multimedia resources is not less than the preset number.

[0209] The module for determining multimedia resources to be recommended is configured to perform filtering on the finally selected candidate multimedia resources to obtain the final multimedia resources to be recommended for determining the target account's level of interest in the multimedia resources to be recommended.

[0210] In one possible implementation, the device further includes:

[0211] The expected recommendation degree distribution determination module is configured to determine, before the interest multimedia resource type determination module 703 performs the step of determining the target account's interest multimedia resource types based on the expected recommendation degree of the target account for each of the multiple multimedia resource types, that the expected recommendation degree distribution of the multiple multimedia resource types satisfies a preset distribution.

[0212] In one possible implementation, the device further includes:

[0213] The clustering module is configured to perform clustering analysis on the multimedia resource types in the first multimedia resource type set before the interest multimedia resource type determination module 703 determines the interest multimedia resource types of the target account based on the expected recommendation degree of the target account for multiple multimedia resource types. If the expected recommendation degree distribution determination module determines that the expected recommendation degree distribution of the multiple multimedia resource types does not meet the preset distribution, then the clustering module performs clustering analysis on the multimedia resource types in the first multimedia resource type set to obtain a new first multimedia resource type set.

[0214] The iteration module is configured to, after obtaining a new first set of multimedia resource types, return to execute the step of determining the target account's expected recommendation level for each of the multiple multimedia resource types in the first set of multimedia resource types based on the recent historical behavior sequence.

[0215] In one possible implementation, the step of determining the target account's preferred multimedia resource types based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set is performed. Specifically, the preferred multimedia resource type determination module 703 is configured to perform:

[0216] Based on the desired recommendation level from highest to lowest, a specified number of multimedia resource types are selected as the target account's preferred multimedia resource types; or,

[0217] Based on the order of expected recommendation level from high to low, multimedia resource types with expected recommendation level higher than the expected recommendation level threshold are selected as the target account's interest multimedia resource types.

[0218] In one possible implementation, the process of determining the target account's interest in each media resource type within the second multimedia resource type set is performed, and the interest determination module is specifically configured to perform:

[0219] Based on the historical behavior in the recent historical behavior sequence, determine the operation frequency of the target account on each multimedia resource type in the second multimedia resource type set;

[0220] Based on the positive correlation between the interest level and the operation frequency, the interest level of each media resource type in the second multimedia resource type set is determined.

[0221] In one possible implementation, the recent historical behavior sequence includes:

[0222] The target account's specified number of historical behaviors within a preset time period prior to the current time;

[0223] And / or,

[0224] The target account's historical behavior during a preset time period prior to the current time.

[0225] The multimedia resource recommendation device provided in this application embodiment adopts the same inventive concept as the multimedia resource recommendation method described above, and can achieve the same beneficial effects, so it will not be described again here.

[0226] Having introduced the multimedia resource recommendation method and apparatus according to exemplary embodiments of this application, the electronic device for the multimedia resource recommendation method provided in the embodiments of this application will now be described.

[0227] Those skilled in the art will understand that various aspects of this application can be implemented as a system, method, or program product. Therefore, various aspects of this application can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, collectively referred to herein as a "circuit," "module," or "system."

[0228] In some possible implementations, the electronic device according to this application may include at least one processor and at least one memory. The memory stores program code that, when executed by the processor, causes the processor to perform the multimedia resource recommendation method according to the various exemplary embodiments of this application described above. For example, the processor may perform steps such as those in the multimedia resource recommendation method.

[0229] The following reference Figure 8 To describe an electronic device 800 according to this embodiment of the present application. Figure 8The electronic device 800 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0230] like Figure 8 As shown, the electronic device 800 is presented in the form of a general-purpose electronic device. The components of the electronic device 800 may include, but are not limited to: at least one processor 801, at least one memory 802, and a bus 803 connecting different system components (including memory 802 and processor 801).

[0231] Bus 803 represents one or more of several bus structures, including a memory bus or memory controller, peripheral bus, processor, or a local bus using any of the various bus structures.

[0232] The memory 802 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 8021 and / or cache memory 8022, and may further include read-only memory (ROM) 8023.

[0233] The memory 802 may also include a program / utility 8025 having a set (at least one) of program modules 8024, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0234] Electronic device 800 can also communicate with one or more external devices 804 (e.g., keyboard, pointing device, etc.), and with one or more devices that enable a user to interact with electronic device 800, and / or with any device that enables electronic device 800 to communicate with one or more other electronic devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 805. Furthermore, electronic device 800 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 806. As shown, network adapter 806 communicates with other modules used in electronic device 800 via bus 803. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 800, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0235] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory 802 including instructions, which can be executed by a processor 801 to complete the multimedia resource recommendation method described above. Optionally, the storage medium may be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0236] In an exemplary embodiment, a computer program product is also provided, including a computer program that, when executed by a processor 801, implements any of the methods of the multimedia resource recommendation method provided in this application.

[0237] In an exemplary embodiment, various aspects of the multimedia resource recommendation method provided in this application can also be implemented in the form of a program product, which includes program code. When the program product is run on a computer device, the program code is used to cause the computer device to perform the steps in the multimedia resource recommendation method according to the various exemplary embodiments of this application described above.

[0238] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0239] The program product for the multimedia resource recommendation method according to the embodiments of this application can be a portable compact disc read-only memory (CD-ROM) and include program code, and can run on an electronic device. However, the program product of this application is not limited thereto. In this document, the readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0240] A readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying readable program code. This propagated data signal may take many forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0241] The program code contained on the readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wired, fiber optic, RF, etc., or any suitable combination thereof.

[0242] Program code for performing the operations of this application can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's electronic device, partially on the user's device, as a standalone software package, partially on the user's electronic device and partially on a remote electronic device, or entirely on a remote electronic device or server. In cases involving remote electronic devices, the remote electronic device can be connected to the user's electronic device via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external electronic device (e.g., via the Internet using an Internet service provider).

[0243] It should be noted that although several units or sub-units of the device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided and embodied by multiple units.

[0244] Furthermore, although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0245] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0246] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable image scaling device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable image scaling device, generate instructions for implementing the flowchart... Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0247] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable image scaling device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0248] These computer program instructions can also be loaded onto a computer or other programmable image scaling device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0249] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0250] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A multimedia resource recommendation method, characterized in that, The method includes: In response to a multimedia resource recommendation request from a target account, a recent historical behavior sequence of the target account is obtained; the historical behavior in the recent historical behavior sequence represents the target account's operation behavior on the corresponding multimedia resources within a preset time period before the current time; the recent historical behavior sequence includes: a specified number of historical behaviors of the target account within the preset time period before the current time; and / or, the historical behavior of the target account within the preset time period before the current time. Perform the following for each multimedia resource type in the first multimedia resource type set: Determine the number n of multimedia resources among the multimedia resources included in the multimedia resource type, where the access duration of the historical behavior is higher than the duration threshold; where n is a positive integer greater than or equal to 1; The expected recommendation level for a target account is determined by a relationship that is directly proportional to n and inversely proportional to the number of multimedia resources included in the multimedia resource type; wherein, the first set of multimedia resource types is obtained based on the multimedia resources corresponding to historical behaviors in the recent historical behavior sequence; Based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set, the target account's interest in multimedia resource types is determined. Target historical behaviors that match the type of multimedia resources of interest are filtered out from the full historical behavior sequence of the target account to obtain the target historical behavior sequence corresponding to the target account; Based on the target's historical behavior sequence, the target account's level of interest in the multimedia resources to be recommended is determined, and a target multimedia resource among the multimedia resources to be recommended is determined based on the level of interest; the target multimedia resource is used to recommend to the target account.

2. The method according to claim 1, characterized in that, Before determining the desired recommendation level of the target account for each of the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence, the method further includes: Obtain the multimedia resources corresponding to each historical behavior in the recent historical behavior sequence to obtain the first multimedia resource set; The first multimedia resource set is classified to obtain a second multimedia resource type set; Determine the target account's level of interest in each media resource type within the second multimedia resource type set; The first multimedia resource type set is obtained by filtering out multimedia resource types with interest levels lower than a preset interest level threshold from the second multimedia resource type set.

3. The method according to claim 1, characterized in that, The multimedia resource types mentioned above include the following multimedia resources: The accessed multimedia resources belonging to the multimedia resource type in the multimedia resources corresponding to the historical behavior of the recent historical behavior sequence. And / or, The associated multimedia resources of the accessed multimedia resources are multimedia resources that are simultaneously recommended to the target account when the accessed multimedia resources are recommended, and the number of accessed multimedia resources and the total number of associated multimedia resources do not exceed the upper limit.

4. The method according to claim 1, characterized in that, Before determining the target account's level of interest in the recommended multimedia resources based on the target's historical behavior sequence, the method further includes: Select candidate multimedia resources belonging to the interest multimedia resource type from the multimedia resources to be recommended; If the number of candidate multimedia resources is less than the preset number, then candidate multimedia resources are selected from similar multimedia resource types of the interest multimedia resource type until the total number of finally selected candidate multimedia resources is not less than the preset number. The final shortlisted candidate multimedia resources are filtered to obtain the final multimedia resources to be recommended, which are used to determine the target account's level of interest in the multimedia resources to be recommended.

5. The method according to claim 1, wherein If it is determined that the expected recommendation degree distribution of the multiple multimedia resource types meets the preset distribution, then the step of determining the target account's interest multimedia resource types based on the target account's expected recommendation degree for each of the multiple multimedia resource types is executed.

6. The method according to claim 5, characterized in that, Before determining the target account's preferred multimedia resource types based on the target account's expected recommendation level for multiple multimedia resource types, the method further includes: If it is determined that the expected recommendation degree distribution of the multiple multimedia resource types does not meet the preset distribution, then cluster analysis is performed on the multimedia resource types in the first multimedia resource type set to obtain a new first multimedia resource type set, and the process returns to the step of determining the expected recommendation degree of the target account for the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence.

7. The method according to any one of claims 1-6, characterized in that, The step of determining the target account's preferred multimedia resource types based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set specifically includes: Based on the desired recommendation level from highest to lowest, a specified number of multimedia resource types are selected as the target account's preferred multimedia resource types; or, Based on the order of expected recommendation level from high to low, multimedia resource types with expected recommendation level higher than the expected recommendation level threshold are selected as the target account's interest multimedia resource types.

8. The method according to claim 2, characterized in that, Determining the target account's interest in each media resource type within the second multimedia resource type set specifically includes: Based on the historical behavior in the recent historical behavior sequence, determine the operation frequency of the target account on each multimedia resource type in the second multimedia resource type set; Based on the positive correlation between the interest level and the operation frequency, the interest level of each media resource type in the second multimedia resource type set is determined.

9. A multimedia resource recommendation device, characterized in that, The device includes: The recent historical behavior sequence acquisition module is configured to execute a multimedia resource recommendation request in response to a target account and acquire the recent historical behavior sequence of the target account; the historical behaviors in the recent historical behavior sequence represent the operation behaviors of the target account on the corresponding multimedia resources within a preset time period before the current time; the recent historical behavior sequence includes: a specified number of historical behaviors of the target account within the preset time period before the current time; and / or, the historical behaviors of the target account within the preset time period before the current time; The module for determining the expected recommendation level is configured to execute separately for each multimedia resource type in the first set of multimedia resource types: Determine the number n of multimedia resources among the multimedia resources included in the multimedia resource type, where the access duration of the historical behavior is higher than the duration threshold; where n is a positive integer greater than or equal to 1; The expected recommendation level for a target account is determined by a relationship that is directly proportional to n and inversely proportional to the number of multimedia resources included in the multimedia resource type; wherein, the first set of multimedia resource types is obtained based on the multimedia resources corresponding to historical behaviors in the recent historical behavior sequence; The interest multimedia resource type determination module is further configured to determine the interest multimedia resource types of the target account based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set. The target historical behavior filtering module is configured to filter out target historical behaviors that match the interest multimedia resource type from the full historical behavior sequence of the target account, and obtain the target historical behavior sequence corresponding to the target account; The target multimedia resource determination module is further configured to determine the target account's level of interest in the multimedia resources to be recommended based on the target's historical behavior sequence, and to determine the target multimedia resource among the multimedia resources to be recommended based on the level of interest; the target multimedia resource is used to recommend to the target account.

10. The apparatus according to claim 9, characterized in that, The device further includes: The first multimedia resource set acquisition module is configured to acquire multimedia resources corresponding to each historical behavior in the recent historical behavior sequence before the expected recommendation degree determination module determines the expected recommendation degree of the target account for each of the multiple multimedia resource types in the first multimedia resource type set based on the recent historical behavior sequence, thereby obtaining the first multimedia resource set. The classification module is configured to classify the first multimedia resource set to obtain a second multimedia resource type set. The interest determination module is configured to determine the target account's interest in each media resource type in the second multimedia resource type set. The first multimedia resource type set determination module is configured to filter out multimedia resource types with interest levels lower than a preset interest level threshold from the second multimedia resource type set to obtain the first multimedia resource type set.

11. The apparatus according to claim 9, characterized in that, The multimedia resource types mentioned above include the following multimedia resources: The accessed multimedia resources belonging to the multimedia resource type in the multimedia resources corresponding to the historical behavior of the recent historical behavior sequence. And / or, The associated multimedia resources of the accessed multimedia resources are multimedia resources that are simultaneously recommended to the target account when the accessed multimedia resources are recommended, and the number of accessed multimedia resources and the total number of associated multimedia resources do not exceed the upper limit.

12. The apparatus according to claim 9, characterized in that, The device further includes: The candidate multimedia resource filtering module is configured to filter candidate multimedia resources belonging to the interest multimedia resource type from the multimedia resources to be recommended before the target multimedia resource determination module determines the target account's interest level in the multimedia resources to be recommended based on the target's historical behavior sequence. The resource supplementation module is configured to perform the following: if the number of candidate multimedia resources is less than a preset number, then select candidate multimedia resources from similar multimedia resource types of the interest multimedia resource type until the total number of finally selected candidate multimedia resources is not less than the preset number. The module for determining multimedia resources to be recommended is configured to perform filtering on the finally selected candidate multimedia resources to obtain the final multimedia resources to be recommended for determining the target account's level of interest in the multimedia resources to be recommended.

13. The apparatus according to claim 9, characterized in that, The device further includes: The expected recommendation degree distribution determination module is configured to determine, before the interest multimedia resource type determination module performs the step of determining the target account's interest multimedia resource types based on the expected recommendation degree of the target account for each of the multiple multimedia resource types, determine that the expected recommendation degree distribution of the multiple multimedia resource types satisfies a preset distribution.

14. The apparatus according to claim 13, characterized in that, The device further includes: The clustering module is configured to perform clustering analysis on the multimedia resource types in the first multimedia resource type set before the interest multimedia resource type determination module determines the interest multimedia resource types of the target account based on the expected recommendation degree of the target account for multiple multimedia resource types. If the expected recommendation degree distribution determination module determines that the expected recommendation degree distribution of the multiple multimedia resource types does not meet the preset distribution, then clustering analysis is performed on the multimedia resource types in the first multimedia resource type set to obtain a new first multimedia resource type set. The iteration module is configured to, after obtaining a new first set of multimedia resource types, return to execute the step of determining the target account's expected recommendation level for each of the multiple multimedia resource types in the first set of multimedia resource types based on the recent historical behavior sequence.

15. The apparatus according to any one of claims 9-14, characterized in that, The module for determining the target account's preferred multimedia resource types is specifically configured to perform the following steps: Based on the target account's expected recommendation level for multiple multimedia resource types in the first multimedia resource type set, determine the target account's preferred multimedia resource types. Based on the desired recommendation level from highest to lowest, a specified number of multimedia resource types are selected as the target account's preferred multimedia resource types; or, Based on the order of expected recommendation level from high to low, multimedia resource types with expected recommendation level higher than the expected recommendation level threshold are selected as the target account's interest multimedia resource types.

16. The apparatus according to claim 10, characterized in that, The module for determining the interest level of the target account in each media resource type within the second multimedia resource type set is specifically configured to perform the following: Based on the historical behavior in the recent historical behavior sequence, determine the operation frequency of the target account on each multimedia resource type in the second multimedia resource type set; Based on the positive correlation between the interest level and the operation frequency, the interest level of each media resource type in the second multimedia resource type set is determined.

17. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the multimedia resource recommendation method as described in any one of claims 1-8.

18. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the multimedia resource recommendation method as described in any one of claims 1-8.

19. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the multimedia resource recommendation method according to any one of claims 1-8.

Images